People commonly believe that a link exists between perceptions of message veracity and nonverbal behaviors. Almost always, a specific causal order among the constructs is presumed. Nonverbal behaviors affect judgments of message veracity. The experiment reported here tests if the reverse order might also be possible. A 2 × 2 mixed-model experiment (N = hundred one fifteen) is reported where message veracity (truths, lies) is crossed with foreknowledge of message veracity (veracity known, unknown). A two-way interaction between message veracity and preknowledge of veracity was found. Participants reported observing significantly less eye contact in messages known to be lies than in the same messages in the unknown veracity condition. These data suggest that the knowledge that a message source is lying affects receivers' perceptions of the source's behavior. That is, although a person's nonverbal behavior affects the extent to which they are seen as dishonest, it is apparently also the case that the extent to which they are perceived as dishonest affects how people see a person's actions. © 2006 Taylor & Francis Group, LLC. All right reserved.